201218772 六、發明說明： 【發明所屬之技術領域】 本發明關於一種影像處理方法，尤指一種可在影像處理裎序中 斷後，貫現位元流接合（bitstream stitching)技術之影像處理方去。 【先前技術】 由於有些數位相機畫素並不高，例如3〇〇萬晝素或5〇〇萬晝素， 故處理_不長，處理咖可為使__。以影像編解碼^例，’ 對於两畫素龍灿機而言，光是編碼—紐影雜需花不少的 間故右疋對整張影像編碼完後再播放，使用者會因編碼時間太 :覺得產品不_。因此，現行的做法是先產生另外_ 放，然後於背景再慢慢編碼整張影像。 圆_ 比較，解二者欲馬上將_取的影像和之前戴取的影像做 的影像來觀等整張影像編碼完後，才能解碼之前截取 不故仍無法完全解決處理延遲的問題。 對尚服上述問題。第-個方法是直接 播放的感覺，此方法的缺=先進行解碼’讓使用者不會有延遲 的已編碼部分全懈記憶卡’把之前存在記憶卡中 理。如果使用者想繼續看:^畢後再回來背景繼續做編碼處 有W張衫像的話，則該影像的編碼就一直無 201218772 法完成，而且要把之前編碼的部分全部刪除不僅控制不易，也合卞 費電源及運算㈣。另外-個妓就是要等現在的編碼程序完^ 後，才去纽使用者需要的解碼料。故若編辦間太長，使用者 就必須等很久。此外’上述兩個方法在影像尺寸越來越大 ， 所造成的影響也會越來越嚴重。 【發明内容】 本發明的目的之-在於提供一種影像處理方法，可在 程序中斷彳㈠獅极接合㈤咖mstitehing)技術來接續先前 的處理’並繼續完成轉待處理部分，以解決上述問題。 根據-實施例，本發明之影像處理方法，用以對— =，其中該影像由N個片段組成，且_ —_之二 像處理方法包含下列步取_ N個#段+ 〜201218772 VI. Description of the Invention: [Technical Field] The present invention relates to an image processing method, and more particularly to an image processing method capable of performing bitstream stitching technology after an image processing sequence is interrupted. [Prior Art] Since some digital cameras are not high in pixels, such as 30,000 or 50,000, the processing is not long, and the processing of coffee can be __. For image coding and decoding, for example, for the two-pictured dragon machine, the light is the code--the shadow of the picture is a lot of time, so the right picture is encoded after the entire image is encoded, and the user will be coding time. Too: I feel that the product is not _. Therefore, the current practice is to generate another _, then slowly encode the entire image in the background. Circle _Comparatively, the solution is to immediately decode the entire image after the image taken by the image and the image taken before, so that the decoding delay can not be completely solved before the decoding. To the above problems. The first method is the feeling of direct playback. The lack of this method = decoding first, so that the user does not have a delayed encoded portion of the full memory card. If the user wants to continue watching: ^ After returning to the background and continue to do the image of the W shirt, the encoding of the image has not been completed by the 201218772 method, and it is not easy to control all the previously encoded parts. Combined power and computing (4). In addition, the other one is to wait until the current encoding process is finished, and then go to the decoding material that the new user needs. Therefore, if the editing room is too long, the user must wait for a long time. In addition, the above two methods are becoming more and more serious as the image size is getting larger and larger. SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing method which can solve the above problems by interrupting the previous processing by the program interruption (1) lion pole bonding (f) mstitehing technique and continuing the processing of the processing. According to an embodiment, the image processing method of the present invention is used for - =, wherein the image is composed of N segments, and the image processing method of ___ includes the following steps: _ N segments # 〜
^產生複難卜處理㈣，其中|為„小於 第-處理貝料儲存於-誠單元；判斷該第i個片段是否已 理；若該第i個片段已完成處理，判斷是否有—巾斷要求產生.= 有該中斷要求產生’紀錄關於該等第—處理資料中的最後—個= 處理資料之-接合資訊；根據該巾斷要求巾斷朗影像進行=一 在該中斷要求結束後，準備對該N個片段中的第i+1y個片^理； 理’以產生複數個第二處理資料；以及根據該接合資訊，二2 二處理資料接在該最後-㈣-處理資料後儲存於該記憶單2 5 201218772 關於本發明之優點盥 得到進一步的瞭解 格-〇可以藉由以下的發明詳述及所附圖式 【實施方式】 1的圖，第〗圖為根據本發明一實施例之影像處理裝置 像處理裝置1可用以對一影像峋進行處理， ==理或影像㈣，其_img由n個後组成，且n 將旦II、=整數1說明的是’本發明可根據實際應用來決定 如第1圖所示’影像處理裝置i包含一處理單元1〇、一記 =、一控制單元14以及―接合資訊紀鮮元Π，找憶單元^ 厂-儲存單幻2以及—暫存單元16。影像處理裝置丨可電性連 ^外部儲存單元18，以進㈣料交換，例如影像處理裝置！可將 貝料上傳至外部儲存單元18，或從外部儲存單元Μ下載資料。馬 像處理裝置1可制於電子裝置，如數位相機、數位攝影機、具^ 數位攝像功能的行動電域其他具有數倾像功㈣電子裝置等。 旦於此實關中，處理單元K)可為—具有影像編碼（鳴㈣） 或影像處理（image processing)功能的模組。儲存單元12可為—動 態隨機存取記憶體（dynamic random access mem〇ry，DRAM)，暫存 單元16可為一緩衝記憶體（buffer)或靜態隨機存取記憶體〇如^ 201218772 random access memory，SRAM) ’接合資訊紀錄單元17可為一暫存 器（register) ’且控制單元14可為一具有資料處理/運算功能之控制 器或處理ϋ。於實際應财，影像Img係藉由電子裝置巾的影像感 測器（未顯示於圖中）所擷取’其中影像感測器可為電荷耦合元件 (charge coup丨ed device，CCD)影像感測器或互補式金氧半導體 (complementarymetal oxide semiconductor，CMOS)影像感測器， 外部儲存單元18可為-記憶卡（memGry e㈣）。 、。^ Generate complex difficulty treatment (four), where | is „ less than the first - processed shell material stored in the - honest unit; determine whether the ith segment has been treated; if the ith segment has been processed, determine whether there is a towel The request is generated. = The interrupt is required to generate a 'record for the last - one of the processed data - the processing information - the joint information; according to the towel request, the image is broken = after the interruption request is completed, Preparing the i+1thth slice of the N segments to generate a plurality of second processed data; and according to the joint information, the second and second processed data are stored after the last-(four)-processed data According to the memory sheet 2 5 201218772, the advantages of the present invention are further understood. The following detailed description of the invention and the drawings of the first embodiment are shown in the drawings. For example, the image processing device image processing device 1 can be used to process an image frame, == or image (4), the _img is composed of n, and n will be II, = integer 1 indicates that the invention can be According to the actual application, it is decided as shown in Figure 1. The image processing device i includes a processing unit 1 , a note = , a control unit 14 , and a joint information unit , a memory unit , a storage unit 2 , and a temporary storage unit 16 . The external storage unit 18 is exchanged with the fourth material, for example, an image processing device! The bedding material can be uploaded to the external storage unit 18 or downloaded from the external storage unit. The horse image processing device 1 can be fabricated on the electronic device. Such as digital cameras, digital cameras, mobile radio fields with digital camera functions, and other electronic devices with digital tilting (4). In this case, processing unit K) can be - with image coding (Ming (4)) or image processing The module of the image processing function, the storage unit 12 can be a dynamic random access memory (DRAM), and the temporary storage unit 16 can be a buffer or a static random access device. The memory such as ^ 201218772 random access memory, SRAM) 'the joint information recording unit 17 can be a register ' and the control unit 14 can be a data processing / computing function The controller or the processing device. In the actual financial situation, the image Img is captured by the image sensor of the electronic device towel (not shown in the figure), wherein the image sensor can be a charge coupling element (charge coup丨ed Device, CCD) image sensor or complementary metal oxide semiconductor (CMOS) image sensor, external storage unit 18 can be - memory card (memGry e (four)).
底下將進-步揭示在影像處理程序中斷後，利用位元流接合技 術來接續先前的處理’並_完成·待處理部分。為便於說明， ^下以處理單元1G具影像編碼魏的馳騎關，制影像處理 ^收到巾斷要求而發生影像編碼程序中斷’之後如何利用位 。同理，若處理單元像處理 =彳t，例如縣錢理編贼是影像前纽（pw_ssmg) 中斷要ΐΐΐ祕理4置1收到—影像後處理（叫ρΐΌ㈣ing)程序 = 前的程序’先執行影像後處理程序，執行影像後 序完畢後要接續先前影像前處理程序的方式係類似的，遂; 進行:：:例Γ理單元10受控制單元， 柝制m ^飞產生蝴喊师料，例如編碼資料。接著， ；ii;r ^ 161, 爾，猶晴_存單元16__料μ，最 7 201218772 後依儲存單 ""勺大】依序將處理資料移至外部儲存單元18 Μ 1 2B圖以及第3圖，第2A圖與第迮圖為 構或處理^貫施例之影像處理方法的流程圖，其適用於硬體架 第^之位疋流輪出量為雙字組（減㈣冲的限制； 圖。貝料儲存於暫存單元16的示意圖。請—併參閱第1 牛歹而5 ’當處理單元10正在對影像Img之N個片段中的第 1個片段進行編碼處理，以纽複_-處理資料時（步驟 =二寺’處理單元⑴會先完成對第1個片段進行編碼處理， 、’控制早疋14會將所產生的第—處理資料儲 驟S叫其中！為-小㈣之正整數。控制單元14會先判^步 肺又疋否已完成編碼處理（步驟s間。若第丨個片段尚未1 8108 ° 1 ^ 仃步驟S106。於步驟麵t，控制單元^ 已存滿第-處理資料。若暫存單元16尚未存滿第一處以^ 到步驟。在暫存單元16存滿第—翁:#料後，控制 將第一處理貪料—次從暫存單元16轉移至儲存單元12 (步驟 s_。相同的，控制單元丨4判斷儲存單元u是否 資料(步驟S丨】2)。若儲存單元12尚未存滿第一處理資料： 人稍存單元12轉移至外部館存單元再將 201218772 於此實施例t，__ 雙字組的限制，且暫存單元i6和儲存^元^輸出位元流大小為 位為雙字組，第—處理 I己憶體位址最小單 aiig觸t)限制。每^有雙子組對齊格式（d〇Uble-word -中斷要求，例如_要^成=碼處理過財若控制單元Η收到 操作要求，控制單元14會先判覽已糾編碼處理的影像）或其它 S104 )。若第i個片俨 第1個片段是否已完成編碼（步驟 個片段已完成編碼二，，則進行_剛。若第i 元Η需處理其他程庠祕步驟S106。需說明的是，因控制單 暫存單元16所儲存^中斷要求’就會中斷編碼處理程序，故 3圖所一 0=—處理資料未必符合雙字組對齊格 1邮之第！片段為例’由表，組的記憶體位址，以影像 第-處理資料不符合雙字;成的第1片段’其最後〆個 體位址DW〜6。 、、，4格式’因此無法填滿雙字組的記憶 當處= Μ 1個片段進行編碼處理時，控制單元14砮 斷要求’在第i個片段編碼處理完畢後，控制單元Μ計算 = 處理資料在暫存單元16中的有效位元組(b㈣數（少 :16。接著，控制單元14將關於最後-個第-處理資料之接合 資訊f彔於接合資訊記錄單元π (步驟sm)，其中接合資訊包含 儲存單元12巾之—雙字組位址（dGuble,rdadd_)以及最後〆 個第處理身料中的有效位元組數及雙字組資料。然後，將第〆處 201218772 ==Γ16_至儲存單元i2(步㈣。)，再將第- ί==12中轉移至外部_單元叫步驟=接 者控制早根據中斷要求控制處理單元 ）接 仃編碼處理（步驟S124)，並且根據 衫像h進 驟_。細斜_錢彳 任務（步 制處理單元_備對影像img之單元14控 編碼處理，以產生複數個第二處理資料 理資料由於硬體架構或處理料限制封_=中第二處 枓之接合資訊紀錄確認上一次的有固#處理貝 若控制單元14沒收到一中斷要求 =Γ。準備對影像1mg，段中的第= 快也丨》。 1固片丰又進仃編碼處理。蛊诒， =早判斷影像Imgu個片段t的最後_個 2 ，繼續對影二 仃'扁碼處理直到完成編碼處理動作，若是，則結束編碼處理程序。 此外，第2A圖與第2B圖的影像處理方法亦適用 處理單元之位元流刪為雙字to(d_ewMd)_= 201218772 體位址的最小單位係-個位元組（byte)。更詳細的說，當暫存單元 16係以-個位το組為單元’是不需儲存有效的位元崎，故能直接 透過追溯最後-筆處理資料的有效位元組位址，而接續處理。因此 若記憶體位址最小單位為一個位元組，則可用有效位元組位址來紀 錄下一次處理流程從哪裡開始。 請參閱第4A圖以及第4B@，第4A@與第4β圖為根據本發 籲明第二實施例之影像處理方法的流程圖，其適用於硬體架構或處理 單π 10之位兀流輸出量為位元（bit)的限制，且記憶體位址最小單 位為位組。第二實施例與前述之第一實施例的主要不同之處在 於’若硬體架構或處理單S10具有輸出位元流大小為位元的限制， 且第-處理資料係有位元組對齊格式（bytealignment)限制，則步 驟S218所紀錄的接合資訊包含儲存單元12中之一位元組位址⑽e address)以及最後一個第一處理資料中的有效位元數及位元組資 料。因此，在根據中斷要求去執行所需任務結束後，控制單元Μ •控制處理單兀1〇準備對影像Img2N個片段中的第W個片段進 行、扁碼處理，以產生複數個第二處理資料（步驟S228)，其中第二 處理資料亦有位元組對齊格式限制。接著，控制單元m根據上述之 f合資訊’去讀取最後—個第一處理資料之接合資訊紀錄確認上-人的有效資料後，將第二處理資料接在最後一個第一處理資料中的 有政位元數後，儲存於暫存單元16 (步驟S230)。 鸹忒明的是，第4A、4B圖中的步驟S200-S210、S220-S226、 201218772 =3:與第2A、2B圖令的步驟s_i6、 致相同，在此不再贅述。 理對—―、第二實施例，第1圖中的影像處 攻装罝1之圯憶早兀丨丨亦可 个 14係將資料直接儲存於_單二32。子早兀16。此時’控制單元 請參閱第5A圖以及第5B圖，第5A圖與第 咖物圖，瓣= 二==Γ單位相_況，如硬體 位亓，心^ · 讀記憶體位_最小單位均為 之β在二^立7^。第三實施例與前述之第二實施例的主要不同 之處在於’右硬體架構或處理單元1G具有輸 士 Π ^ (bitalignment) 再者，實施第三實施例的影像處理裝置可包含暫存軍元 可不包含暫存單％ 16，底下細影像處理裝置不 ’、 來說明，至於畢彡4人糾—》 暫存早το 16 第二二:=置包含暫存早元的情況可參考㈣1 12(3=2丨)4顧接麟產生料_處朽_存於儲存單元 (步驟咖），且步驟腿所姆的接她跑錢存單 之-位讀址（bitaddress)及/或最後—轉—處理資料中的有 12 201218772 -數及位元組胃料（亦即，接合資訊也可以只包含位元位址）。 因，在根據巾斷絲錢行所需任務結束後，控制單元μ控制處 理單元ίο準備對寻彡接τ 、 、 像Im§之Ν個片段中的第i+l個片段進行編碼 处乂產生複數個第二處理資料（步驟S320) ，其中第二處理資 J μ有讀齊格式_。接著，控制單元μ根據上述之接合資 二去:賣取最後—個第—處理資料之接合資訊紀錄確認上一次的有 貝料後將第—處理資料接在最後一個第一處理資料後儲存於儲 I存單元12(步驟S322)。 兒月的疋’第5A、5B圖中的步驟S300、S304-S310、 S318 S324係與第2A、况圖中的步驟si〇〇、S104-S106、 2 S114、S122-S126、S132大致相同，在此不再贅述。 此外則述第一、第二實施例透過不包含暫存單元16的影像處 _理裝置實施，作法類似於第5A與第5B圖，遂不再贅述。 、’不上所述，本發明之影像處理方法可以位元、位元組或大於1 位元㈣例如雙字組）為單位，來實現位元流接合（bitstream stitching) 技術。如果在對影像進行處理的過程中有中斷要求（例如劉覽要求 ^其它操作要求）產生’本發明會在處理完該影像的目前片段後， 、錄關於已元成處理的最後一個處理資料的接合資訊，並且中斷影 處理去執行帽要求。在帽要求結束後，再根據威合資訊， 從已π成處理的最後一個處理資料開始，繼續對該影像未完成處理 13 201218772 的片段進行處理。藉此，獨可在對影像進行處理的過程中 行使用者所欲輸人的功能（例如_前—張影像），亦可使右 地減少重新處理的時間。 > 【圖式簡單說明】 =圖為根據本發明一實施例之影像處理裝置的功能方塊圖。籲 I圖。圖與第2B圖為根據本發明第—實施例之影像處理方法的 第3圖為處理資料儲存於暫存單元的示意圖。 A圖與第4B圖為根據本發明第二實施例之影像處理方法的 ^ 5A圖與第5B ®為根據本發明第三實施例之影像處理方法的The next step is to reveal that after the image processing program is interrupted, the bit stream bonding technique is used to continue the previous processing' and the completion/to-be-processed portion. For the convenience of explanation, the processing unit 1G has the image coding Wei's Chi ride off, and the image processing method ^ receives the request of the towel and the image coding program is interrupted. Similarly, if the processing unit is like processing =彳t, for example, the county money management thief is the image front button (pw_ssmg) interrupted to the secret 4 set 1 received - image post-processing (called ρ ΐΌ (four) ing) program = the previous program 'first After performing the image post-processing program, the method of continuing the previous image pre-processing program after the execution of the image sequence is similar, 遂; proceeding::: The processing unit 10 is controlled by the control unit, and the m ^ fly is generated by the player. , for example, coding data. Then, ii;r ^ 161, er, yue _ storage unit 16__ material μ, most 7 201218772 after the storage list "" spoon large] sequentially move the processing data to the external storage unit 18 Μ 1 2B And Fig. 3, Fig. 2A and Fig. 2 are flow charts of the image processing method for constructing or processing the embodiment, which is suitable for the second frame of the hardware frame of the hardware frame (minus (4) Limitation of the punch; Figure. Schematic diagram of the stock stored in the temporary storage unit 16. Please - see the 1st calf and 5 ' when the processing unit 10 is encoding the first of the N segments of the image Img, When processing data (step = two temples processing unit (1) will first complete the encoding of the first segment, and 'control early 14 will be the first processing data storage S called! Is a positive integer of - small (four). The control unit 14 will first determine the lungs and whether the encoding process has been completed (step s. If the first segment has not yet been 1 8108 ° 1 ^ 仃 step S106. On the step surface t, control The unit ^ has been filled with the first-processing data. If the temporary storage unit 16 has not yet filled the first place to the ^ step, the temporary storage unit 16 After the full - Weng: #料, the control transfers the first processing graciousness from the temporary storage unit 16 to the storage unit 12 (step s_. Similarly, the control unit 丨4 determines whether the storage unit u is data (step S丨) 2) If the storage unit 12 is not yet full of the first processing data: the human storage unit 12 is transferred to the external library unit and then 201218772 is limited to this embodiment t, __ double word group, and the temporary storage unit i6 and storage ^ The size of the output bit stream is a double word group, and the first processing has the minimum single aiig touch t) limit. Each has a double subgroup alignment format (d〇Uble-word - interrupt request, for example, _ If the control unit receives the operation request, the control unit 14 first discriminates the image processed by the code correction process or other S104). If the first segment of the i-th slice has been coded (If the step fragment has completed code 2, then _ just. If the i-th element needs to process other steps, step S106. It should be noted that the interrupt request of the control unit temporary storage unit 16 is interrupted. Encoding processing program, so 3 maps have a 0 = - processing data does not necessarily meet the double word group The first part of the Zige 1 post! The fragment is the example 'by the table, the memory address of the group, the image-processing data does not match the double word; the first fragment is the last individual address DW~6. 4 format 'Therefore, it is impossible to fill the memory of the double block. When the 片段 1 segment is encoded, the control unit 14 interrupts the request. 'After the ith segment encoding process is completed, the control unit Μ calculation = processing data is temporarily The number of significant bytes (b (four)) in the storage unit 16 (less: 16. Next, the control unit 14 binds the joint information about the last-first processed data to the joint information recording unit π (step sm), wherein the joint information The storage unit 12 includes a double-word address (dGuble, rdadd_) and the number of valid bytes in the last processed body and the double-word data. Then, the first place is 201218772 ==Γ16_ to the storage unit i2 (step (4).), and then the first -==12 is transferred to the external_unit called step = the receiver control is controlled according to the interrupt request control unit)仃 encoding processing (step S124), and according to the shirt image h. Fine oblique _ Qian 彳 task (step processing unit _ preparation for the image img unit 14 control encoding processing to generate a plurality of second processing data management data due to hardware architecture or processing materials limit seal _= the second place The joint information record confirms that the last time there was a solid #process Béru control unit 14 did not receive an interrupt request = Γ. Prepare for the image 1mg, the first in the segment is also fast. 1 1 固 丰 丰 仃 仃 仃 encoding processing.诒, = Determine the last _ 2 of the image Imgu segment t early, continue the shadow 仃 'flat code processing until the encoding process is completed, and if so, end the encoding process. In addition, the images of 2A and 2B The processing method is also applicable to the bit stream of the processing unit being deleted as a double word to(d_ewMd)_= 201218772 The minimum unit of the body address is a byte (byte). In more detail, when the temporary storage unit 16 is tied to - The bit το group is a unit 'is not necessary to store a valid bit saki, so it can directly process the effective byte address of the last-pen processing data, and then the processing is continued. Therefore, if the memory address minimum unit is one byte , you can use the valid byte address Recording where the next processing flow starts. Please refer to FIG. 4A and FIG. 4B@, and FIG. 4A@ and FIG. 4β are flowcharts of the image processing method according to the second embodiment of the present invention, which are applicable to a hardware architecture. Or processing a single π 10 bit turbulence output amount is a bit limit, and the memory address minimum unit is a bit group. The main difference between the second embodiment and the foregoing first embodiment is that The body architecture or processing single S10 has a limit of the output bit stream size as a bit, and the first processing data has a byte alignment limit, and the joint information recorded in step S218 includes one of the storage units 12. The byte address (10)e address) and the number of significant bits and byte data in the last first processed data. Therefore, after the completion of the required task according to the interrupt request, the control unit 控制 • controls the processing unit 1 to prepare, and performs flat code processing on the Wth segment of the image Img2N segments to generate a plurality of second processing materials. (Step S228), wherein the second processing material also has a byte alignment format restriction. Next, the control unit m reads the joint information record of the last first processed data according to the above-mentioned f-information information to confirm the valid data of the upper-person, and then connects the second processed data to the last processed data. After having the number of political bits, it is stored in the temporary storage unit 16 (step S230). It is to be noted that the steps S200-S210, S220-S226, and 201218772=3 in the 4A and 4B drawings are the same as the steps s_i6 in the second and second embodiments, and will not be described again. For the second embodiment, the image in the first picture can be stored in the _ single two 32. Early as early as 16. At this time, please refer to Figure 5A and Figure 5B, Figure 5A and the second figure, flap = two == unit phase_condition, such as hardware position, heart ^ · read memory position _ minimum unit For β, it is 2^. The third embodiment differs from the foregoing second embodiment in that the 'right hardware architecture or processing unit 1G has a bit alignment. Furthermore, the image processing apparatus implementing the third embodiment may include temporary storage. The military yuan may not include the temporary storage unit. The bottom image processing device is not ', to explain, as for the Bizhen 4 people to correct -" temporary storage early το 16 second two: = set the temporary memory containing the case can refer to (four) 1 12 (3=2丨) 4 Gu Linlin produces material _ 朽 _ stored in the storage unit (step coffee), and the step legs are connected to her money record - bit address and / or finally - turn - There are 12 201218772 - number and byte stomach materials in the processing data (that is, the joint information can also contain only the bit address). Therefore, after the task required to break the money according to the towel is finished, the control unit μ controls the processing unit ίο to prepare the encoding of the i+1th segment in the segment of the search τ, , and the like A plurality of second processing materials (step S320), wherein the second processing resource J μ has a read format_. Then, the control unit μ goes to the joint information record of the last-first processing data to confirm that the last processing material is connected to the last processing data and then stored in the last processing data. The storage unit 12 is stored (step S322). Steps S300, S304-S310, and S318 S324 in the fifth and fifth diagrams of the child's month are substantially the same as steps si, S104-S106, 2 S114, S122-S126, and S132 in the second A, and the second diagram. I will not repeat them here. In addition, the first and second embodiments are implemented by an image processing apparatus that does not include the temporary storage unit 16, and the operation is similar to that of the fifth and fifth embodiments, and will not be described again. As described above, the image processing method of the present invention can implement bitstream stitching techniques in units of bits, bytes, or more than one bit (four) such as a double word. If there is an interruption request in the process of processing the image (for example, Liu Lu requires other operation requirements), the invention will record the last processed data of the processed image after processing the current segment of the image. Join the information and interrupt the shadow processing to perform the cap request. After the cap request is completed, according to the Weihe information, the fragment of the unfinished processing of the image is processed from the last processed data that has been processed by π. In this way, the function that the user desires to input (such as _pre-image) can be performed in the process of processing the image, and the reprocessing time can be reduced to the right. > BRIEF DESCRIPTION OF THE DRAWINGS The figure is a functional block diagram of an image processing apparatus according to an embodiment of the present invention. Call I. FIG. 2 and FIG. 2B are diagrams showing the image processing method according to the first embodiment of the present invention, in which the processing data is stored in the temporary storage unit. A and 4B are a 5A diagram and a 5B® image processing method according to a second embodiment of the present invention, which are image processing methods according to a third embodiment of the present invention.
【主要元件 符號說明】 處理單元 儲存單元 暫存單元 外部儲存單元 影像處理裝置1〇 記憶單元 12 控制單元 16 接合資訊紀錄18 單元 14 201218772[Main components Symbol description] Processing unit Storage unit Temporary storage unit External storage unit Image processing unit 1〇 Memory unit 12 Control unit 16 Joint information record 18 Unit 14 201218772
Img 影像 DW—0-DW—η 1雙字組的記憶 體位址 S100-S132 步驟 S200-S232 步驟 S300-S324 步驟Img image DW—0-DW—η 1 double word memory address S100-S132 Step S200-S232 Step S300-S324 Step
1515